An Implementation of a Database Management System for Real-Time Large-Lake Observations

Abstract

Real-time environmental observations have typically been stored in relatively inaccessible flat text files. In this publication, we present an instance of a PostgreSQL database management system to ingest real-time observations from four buoys in North America's western Lake Erie. Data are transmitted via a cellular data modem, initially archived as text data, and then ingested into the database for further analysis and retrieval. The database utilizes two tables with parallel structure to archive data in a consistent manner. We assign unique keys to instrumentation configurations as they change within and between monitoring seasons. Daily sets of data are linked to their configurations by key and thus are documented, allowing for efficient browsing of user desired data. In addition, we quality check data and archive the findings in a corresponding matrix. With assistance from server-side processing, we produce a web interface for the database. We hope the design of this database allows for relatively simple deployment in domains other than western Lake Erie.